CN116469534A - Hospital number calling management system and method thereof - Google Patents

Hospital number calling management system and method thereof Download PDF

Info

Publication number
CN116469534A
CN116469534A CN202310430032.4A CN202310430032A CN116469534A CN 116469534 A CN116469534 A CN 116469534A CN 202310430032 A CN202310430032 A CN 202310430032A CN 116469534 A CN116469534 A CN 116469534A
Authority
CN
China
Prior art keywords
training
word
appointment information
feature
dimensional
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310430032.4A
Other languages
Chinese (zh)
Inventor
彭亚文
柴伟
白杨
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
First Medical Center of PLA General Hospital
Original Assignee
First Medical Center of PLA General Hospital
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by First Medical Center of PLA General Hospital filed Critical First Medical Center of PLA General Hospital
Priority to CN202310430032.4A priority Critical patent/CN116469534A/en
Publication of CN116469534A publication Critical patent/CN116469534A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/20ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/30Semantic analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • G06N3/0455Auto-encoder networks; Encoder-decoder networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/0464Convolutional networks [CNN, ConvNet]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/047Probabilistic or stochastic networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/048Activation functions
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • G06N3/084Backpropagation, e.g. using gradient descent
    • GPHYSICS
    • G07CHECKING-DEVICES
    • G07CTIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C11/00Arrangements, systems or apparatus for checking, e.g. the occurrence of a condition, not provided for elsewhere
    • GPHYSICS
    • G07CHECKING-DEVICES
    • G07CTIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C11/00Arrangements, systems or apparatus for checking, e.g. the occurrence of a condition, not provided for elsewhere
    • G07C2011/04Arrangements, systems or apparatus for checking, e.g. the occurrence of a condition, not provided for elsewhere related to queuing systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Business, Economics & Management (AREA)
  • Business, Economics & Management (AREA)
  • Medical Informatics (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Public Health (AREA)
  • Probability & Statistics with Applications (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

A hospital number calling management system and a method thereof acquire appointment information of a plurality of objects to be treated arranged in front of a target object to be treated; and mining the correlation characteristic distribution information among the semantic understanding characteristics of the appointment information of each object to be diagnosed by adopting an artificial intelligence technology based on deep learning, accurately estimating the waiting time of the appointment number based on the correlation characteristic distribution information, reasonably planning the own journey based on the estimated waiting time, and improving the experience degree and satisfaction of patients.

Description

Hospital number calling management system and method thereof
Technical Field
The present application relates to the field of intelligent management technologies, and more particularly, to a hospital number calling management system and a method thereof.
Background
With the rapid development of medical systems, the rapid growth of hospital scale and the increasing number of medical persons, the problems of long queuing time for registering and paying, difficult reservation of famous medical specialists, inconvenient medical procedures and the like are particularly outstanding, and the demands of doctors and patients on improving medical environments are becoming stronger.
At present, the registration queuing waiting time is long in the medical treatment process of the patient. In the process of waiting for a patient to call, the traditional reminding waiting time system can only prompt a plurality of people in front, but no method is provided for giving the estimated time, so that the doctor cannot make decisions according to the estimated time, and the experience and satisfaction of the patient are lower.
Accordingly, an optimized hospital call management system is desired.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a hospital number calling management system and a hospital number calling management method, which acquire appointment information of a plurality of objects to be treated, which are arranged in front of a target object to be treated; and mining the correlation characteristic distribution information among the semantic understanding characteristics of the appointment information of each object to be diagnosed by adopting an artificial intelligence technology based on deep learning, accurately estimating the waiting time of the appointment number based on the correlation characteristic distribution information, reasonably planning the own journey based on the estimated waiting time, and improving the experience degree and satisfaction of patients.
In a first aspect, a hospital number calling management system is provided, which includes:
the system comprises a diagnosis appointment information data acquisition module, a diagnosis appointment information acquisition module and a diagnosis appointment information processing module, wherein the diagnosis appointment information acquisition module is used for acquiring diagnosis appointment information of a plurality of objects to be diagnosed arranged in front of a target object to be diagnosed;
the appointment information semantic understanding module is used for enabling the appointment information of each object to be diagnosed to pass through a context encoder comprising a word embedding layer to obtain a plurality of appointment information semantic understanding feature vectors;
the appointment information semantic association module is used for arranging the appointment information semantic understanding feature vectors for the doctor for a plurality of doctor-seeing to be one-dimensional feature vectors and obtaining appointment information semantic association feature vectors through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model; and
And the waiting time evaluation module is used for enabling the reservation information semantic association feature vector to pass through a decoder to obtain a decoding value, wherein the decoding value is used for representing the estimated waiting time.
In the hospital number calling management system, the reservation information semantic understanding module includes: the word segmentation unit is used for carrying out word segmentation processing on the appointment information of each to-be-diagnosed object so as to convert the appointment information of each to-be-diagnosed object into a word sequence consisting of a plurality of words; an embedded encoding unit, configured to map each word in the word sequence to a word vector using an embedding layer of the context encoder including a word embedding layer to obtain a sequence of word vectors; and a context coding unit, configured to perform global context semantic coding on the sequence of word vectors using the converter of the context encoder including the word embedding layer to obtain the plurality of diagnosis appointment information semantic understanding feature vectors.
In the hospital number calling management system, the context coding unit includes: a query vector construction subunit, configured to perform one-dimensional arrangement on the sequence of word vectors to obtain a global word feature vector; a self-attention subunit, configured to calculate a product between the global word feature vector and a transpose vector of each word vector in the sequence of word vectors to obtain a plurality of self-attention association matrices; the normalization subunit is used for respectively performing normalization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of normalized self-attention correlation matrices; the attention calculating subunit is used for obtaining a plurality of probability values through a Softmax classification function by each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and the attention applying subunit is used for weighting each word vector in the word vector sequence by taking each probability value in the plurality of probability values as a weight so as to obtain the plurality of diagnosis appointment information semantic understanding feature vectors.
In the hospital number calling management system, the reservation information semantic association module includes: the first convolution unit is used for respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and activation processing on the one-dimensional feature vector in forward transfer of layers by using each layer of a first one-dimensional convolution neural network model of the double-branch correlation feature extraction structure to take the output of the last layer of the first one-dimensional convolution neural network model as a first-scale reservation information feature vector, wherein the first one-dimensional convolution neural network model is provided with a one-dimensional convolution kernel of a first scale; a second convolution unit, configured to perform convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix, and activation processing on the one-dimensional feature vector in forward transfer of layers of a second one-dimensional convolution neural network model using the two-branch correlation feature extraction structure, so that an output of a last layer of the second one-dimensional convolution neural network model is a second-scale reservation information feature vector, where the second one-dimensional convolution neural network model has a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; and the cascading unit is used for cascading the first-scale reservation information feature vector and the second-scale reservation information feature vector to obtain the reservation information semantic association feature vector.
The hospital number calling management system further comprises a context encoder comprising a word embedding layer, the double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model, and a training module for training the decoder.
In the hospital number calling management system, the training module includes: the training data acquisition module is used for acquiring training data, wherein the training data comprises training diagnosis reservation information of a plurality of objects to be diagnosed and the estimated real value of the waiting time; the training appointment information semantic understanding module is used for enabling the training appointment information of each object to be diagnosed to pass through the context encoder comprising the word embedding layer to obtain a plurality of training appointment information semantic understanding feature vectors; the training appointment information semantic association module is used for arranging the plurality of training appointment information semantic understanding feature vectors into training one-dimensional feature vectors and obtaining training appointment information semantic association feature vectors through the double-branch association feature extraction structure comprising the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model; the optimization module is used for carrying out feature distribution optimization on the training reservation information semantic association feature vectors so as to obtain optimization training reservation information semantic association feature vectors; the decoding loss module is used for enabling the optimization training reservation information semantic association feature vectors to pass through the decoder so as to obtain decoding loss function values; and a training module for training the context encoder including the word embedding layer, the dual-branch correlation feature extraction structure including the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model, and the decoder based on the decoding loss function value and by back propagation of gradient descent.
In the hospital number calling management system, the optimizing module is used for: performing Geng Beier normal periodic re-parameterization on the training reservation information semantic association feature vector by using the following optimization formula to obtain the optimization reservation information semantic association feature vector; wherein, the optimization formula is:
wherein v is i The characteristic values of all positions of the training reservation information semantic association characteristic vector are represented, mu and sigma are respectively the mean value and the variance of a characteristic value set of all positions of the training reservation information semantic association characteristic vector, log represents a logarithmic function based on 2, arcsin(s) represents an arcsine function, arccos(s) represents an arccosine function, v i ' feature values representing respective positions of the optimization training subscription information semantic association feature vector.
In the hospital number calling management system, the solutionA code loss module comprising: the training decoding unit is used for carrying out decoding regression on the semantic association feature vectors of the optimized training reservation information by using the decoder according to the following formula so as to obtain training decoding values; wherein, the formula is:wherein X is the semantic association feature vector of the optimized training reservation information, Y is the training decoding value, W is a weight matrix,/and a method for optimizing the training reservation information >Representing a matrix multiplication; and a loss function calculation unit configured to calculate, as the decoding loss function value, a variance between the training decoded value and a true value of the estimated wait time in the training data.
In a second aspect, a hospital number calling management method is provided, which includes:
acquiring appointment information of a plurality of objects to be treated, which are arranged in front of a target object to be treated;
the appointment information of each object to be diagnosed is respectively passed through a context encoder comprising a word embedding layer to obtain a plurality of appointment information semantic understanding feature vectors;
arranging the plurality of appointment information semantic understanding feature vectors into one-dimensional feature vectors, and obtaining appointment information semantic association feature vectors through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model; and
and the reservation information semantic association feature vector is passed through a decoder to obtain a decoding value, wherein the decoding value is used for representing the estimated waiting time.
In the hospital number calling management method, the method for obtaining a plurality of diagnosis reservation information semantic understanding feature vectors by passing the diagnosis reservation information of each to-be-diagnosed object through a context encoder including a word embedding layer, includes: performing word segmentation processing on the appointment information of each to-be-diagnosed object so as to convert the appointment information of each to-be-diagnosed object into a word sequence composed of a plurality of words; mapping each word in the word sequence to a word vector using an embedding layer of the context encoder including the word embedding layer to obtain a sequence of word vectors; and performing global-based context semantic coding on the sequence of word vectors using a translator of the context encoder including a word embedding layer to obtain the plurality of appointment information semantic understanding feature vectors.
Compared with the prior art, the hospital number calling management system and the hospital number calling management method provided by the application acquire the appointment information of a plurality of objects to be treated, which are arranged in front of a target object to be treated; and mining the correlation characteristic distribution information among the semantic understanding characteristics of the appointment information of each object to be diagnosed by adopting an artificial intelligence technology based on deep learning, accurately estimating the waiting time of the appointment number based on the correlation characteristic distribution information, reasonably planning the own journey based on the estimated waiting time, and improving the experience degree and satisfaction of patients.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments or the description of the prior art will be briefly introduced below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is an application scenario diagram of a hospital number calling management system according to an embodiment of the present application.
Fig. 2 is a block diagram of a hospital number calling management system according to an embodiment of the present application.
Fig. 3 is a block diagram of the subscription information semantic understanding module in the hospital call management system according to the embodiment of the present application.
Fig. 4 is a block diagram of the context encoding unit in the hospital number calling management system according to the embodiment of the present application.
Fig. 5 is a block diagram of the subscription information semantic association module in the hospital call management system according to an embodiment of the present application.
Fig. 6 is a block diagram of the training module in the hospital call management system according to an embodiment of the present application.
Fig. 7 is a block diagram of the decode loss module in the hospital call management system according to an embodiment of the present application.
Fig. 8 is a flowchart of a hospital number calling management method according to an embodiment of the present application.
Fig. 9 is a schematic diagram of a system architecture of a hospital number calling management method according to an embodiment of the present application.
Detailed Description
The following description of the technical solutions in the embodiments of the present application will be made with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
Unless defined otherwise, all technical and scientific terms used in the examples of this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application.
In the description of the embodiments of the present application, unless otherwise indicated and defined, the term "connected" should be construed broadly, and for example, may be an electrical connection, may be a communication between two elements, may be a direct connection, or may be an indirect connection via an intermediary, and it will be understood by those skilled in the art that the specific meaning of the term may be understood according to the specific circumstances.
It should be noted that, the term "first\second\third" in the embodiments of the present application is merely to distinguish similar objects, and does not represent a specific order for the objects, it is to be understood that "first\second\third" may interchange a specific order or sequence where allowed. It is to be understood that the "first\second\third" distinguishing objects may be interchanged where appropriate such that the embodiments of the present application described herein may be implemented in sequences other than those illustrated or described herein.
As described above, currently, registration queuing is long during a patient's visit. In the process of waiting for a patient to call, the traditional reminding waiting time system can only prompt a plurality of people in front, but no method is provided for giving the estimated time, so that the doctor cannot make decisions according to the estimated time, and the experience and satisfaction of the patient are lower. Accordingly, an optimized hospital call management system is desired.
In one embodiment of the application, the hospital number calling management system has the beneficial effects that: 1. the hospital number calling management system can queue and call numbers for initial patients and re-diagnosed patients according to certain rules and mathematical models and algorithms, and rules are different under different conditions. For example: the importance of the first diagnosis is slightly higher than that of the second diagnosis, and the first diagnosis is considered to not see the anxiety and anxiety of the mind, the second diagnosis is the result, 1 person is returned in the second diagnosis, and meanwhile, 2-3 persons are returned for optimal insertion after reports and the like, so that the total waiting time of the person is shorter; 2. autonomous identification and marking of old patients. For example: the doctor is hung continuously for two years by a patient, the old patient is defaulted by the system twice or three times, or the old patient is marked after the surgical operation, the old patient can be marked specifically, and once the reported color difference is identified, the old patient can go to take a film and then return to review, so that the doctor can conveniently treat the doctor; 3. the number passing module is used for processing the number if the patient does not visit in time on the basis of the number calling, and does not display the information of the patient if the number is returned in advance; 4. and a reminder waiting time module. For example: after reporting in the diagnostic area, the two-dimension codes of the public numbers of the hospital are displayed, and a system for scanning the codes can display a plurality of people who are currently in front of you for treatment. The hospital number management system is updated one by one based on the reminder, for example 6,4,2, etc., since if there are consecutive updates, there are 6 people present and 7 waiting for a patient to feel unfair, considering that there may be a re-diagnosis patient insertion. The patient can arrange himself to wait in the diagnosis area according to the dynamic change, or go to another hall to wait for the home visit of the dry spot and return. Furthermore, aiming at the patient needing to hang a plurality of departments or numbers at the same time, the doctor can be selected to see first and then the other doctor can be selected according to prompt service in time, so that the experience and satisfaction of the patient are improved; 5. treatment after the doctor has seen is divided into several categories: assay, examination, drug delivery and no disposal. For laboratory examination and medicine opening, once a doctor opens an laboratory sheet in the system, the doctor can directly carry out a blood drawing flow chart and video demonstration on the platform in the hand of the patient, or take a picture to carry out the examination flow, take medicine and open the flow index of the hospitalization application sheet. Thus, the patient can know the own treatment process later, the efficiency is improved, doctors do not need to explain and inform repeatedly because of the details, and the time of the doctors is saved.
Accordingly, considering that in the process of registering and queuing for waiting for an actual patient, in order to accurately estimate waiting time, to reasonably plan and decide own journey based on the estimated time, for example, for a patient needing to hang several departments or numbers at the same time, the patient can choose which doctor to see first and then another doctor to improve experience and satisfaction of the patient. That is, since the condition and treatment of the treatment are different for each subject, such as assay, examination, start or no treatment, etc., the time spent by each subject at the time of the treatment is different, which is presented on the appointment information of the subject. However, since each subject to be treated has a different condition, the semantic understanding information in the treatment reservation information is also different, and the treatment reservation semantic understanding characteristics of each subject to be treated have an association relationship with respect to the waiting time, for example, when one patient leaves, the total waiting time needs to be recalculated. Therefore, in this process, the difficulty is how to dig out the related feature distribution information among the semantic understanding features of the appointment information of each to-be-diagnosed object, so as to accurately estimate the waiting time of the appointment number, so as to reasonably plan the own journey based on the estimated waiting time, and promote the experience and satisfaction of the patient.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. The development of deep learning and neural networks provides new solutions and schemes for mining the correlation feature distribution information among semantic understanding features of the appointment information of each to-be-diagnosed object.
Specifically, in the technical scheme of the present application, first, the appointment information of a plurality of objects to be treated arranged in front of the target object to be treated is acquired. Then, considering that the appointment information of each to-be-diagnosed object is composed of each word, and semantic association characteristic information of context exists among each word, encoding each appointment information of each to-be-diagnosed object through a context encoder comprising a word embedding layer so as to extract relevant characteristic information of disorder context semantic understanding about each to-be-diagnosed object in each appointment information of each to-be-diagnosed object, namely hidden characteristic information of the diagnosis time of each to-be-diagnosed object, thereby obtaining a plurality of appointment information semantic understanding characteristic vectors.
Further, it is also considered that there is a correlation between the condition context semantic understanding characteristics of the individual subjects to be treated with respect to latency, that is, the time of treatment of the individual subjects to be treated who are ranked in front of the target subject together determines the time that the target subject needs to wait, and when one or more patients in front of the target subject leave, the overall latency also needs to be recalculated. Therefore, in the technical scheme of the application, after the plurality of appointment information semantic understanding feature vectors are arranged into one-dimensional feature vectors, the appointment information semantic association feature vectors are obtained through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model. Particularly, the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model adopt one-dimensional convolutional kernels with different scales to perform feature mining on the one-dimensional feature vectors so as to extract multi-scale relevance feature distribution information among semantic understanding features of disorder waiting time of each object to be diagnosed, and therefore the reservation information semantic relevance feature vectors are obtained. In this way, the time sequence multi-scale associated characteristic information of the treatment waiting time of each treatment object can be extracted.
And then, carrying out decoding regression on the reservation information semantic association feature vector in a decoder to obtain a decoding value for representing the estimated waiting time. That is, decoding is performed according to the multi-scale correlation characteristics of the treatment waiting time of each to-be-treated object arranged in front of the target treatment object, so that the treatment waiting time of the target treatment object is evaluated, the target treatment object can reasonably plan own journey based on the estimated waiting time, and the experience degree and satisfaction of patients are improved.
In particular, in the technical solution of the present application, in order to improve the feature expression effect of the subscription information semantic association feature vector, it is preferable to directly concatenate the first sub-feature vector and the second sub-feature vector obtained by the dual-branch association feature extraction structure of the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model to obtain the subscription information semantic association feature vector, however, considering that the input of the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model is the direct cascade arrangement of a plurality of diagnosis subscription information semantic understanding feature vectors, this may make the continuity of the overall feature distribution of the subscription information semantic association feature vector worse, thereby affecting the training effect.
Based on this, the applicant of the present application performs a normal periodic re-parameterization on the subscription information semantic association feature vector, for example denoted as V Geng Beier (gummel), to obtain an optimized subscription information semantic association feature vector, for example denoted as V', specifically:
mu and sigma are respectively the eigenvalue sets v i Mean and variance of e V, and V i ′∈V′。
Here, the Geng Beier normal periodic re-parameterization is performed by semantically associating the reservation information with the feature values V of the respective positions of the feature vector V i The random periodic operation mode based on Geng Beier (Gumbel) distribution is converted into angular feature expression of probability distribution, and random periodic distribution is introduced into normal distribution of a feature value set to obtain periodic continuous micro approximation of original feature distribution, so that the dynamic continuous wave capacity of counter propagation is improved through periodic re-parameterization of features, the dynamic applicability of a convolution kernel in a training process is improved, and the influence of the poor continuity of the integral feature distribution of reservation information semantic association feature vectors on training effects is compensated. Therefore, the waiting time for the patient to visit the patient can be accurately estimated, so that the journey of the patient can be reasonably planned based on the estimated waiting time, and the experience and satisfaction of the patient are improved.
Fig. 1 is an application scenario diagram of a hospital number calling management system according to an embodiment of the present application. As shown in fig. 1, in the application scenario, first, the appointment information of a plurality of subjects to be treated arranged in front of the target subject to be treated is acquired (e.g., C as illustrated in fig. 1); the acquired appointment information is then input into a server (e.g., S as illustrated in fig. 1) deployed with a hospital call management algorithm, wherein the server is capable of processing the appointment information based on the hospital call management algorithm to generate a decoded value representing the estimated wait time.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
In one embodiment of the present application, fig. 2 is a block diagram of a hospital number calling management system according to an embodiment of the present application. As shown in fig. 2, a hospital number calling management system 100 according to an embodiment of the present application includes: the appointment information data acquisition module 110 is configured to acquire appointment information of a plurality of objects to be appointment arranged in front of a target object to be appointment; the appointment information semantic understanding module 120 is configured to obtain a plurality of appointment information semantic understanding feature vectors by passing the appointment information of the objects to be treated through a context encoder including a word embedding layer; the appointment information semantic association module 130 is configured to arrange the plurality of appointment information semantic understanding feature vectors for appointment information into one-dimensional feature vectors, and obtain appointment information semantic association feature vectors through a dual-branch association feature extraction structure including a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model; and a latency evaluation module 140, configured to pass the reservation information semantic association feature vector through a decoder to obtain a decoded value, where the decoded value is used to represent the estimated latency.
Specifically, in the embodiment of the present application, the appointment information data acquisition module 110 is configured to acquire appointment information of a plurality of to-be-appointment objects arranged in front of the target to-be-appointment object. As described above, currently, registration queuing is long during a patient's visit. In the process of waiting for a patient to call, the traditional reminding waiting time system can only prompt a plurality of people in front, but no method is provided for giving the estimated time, so that the doctor cannot make decisions according to the estimated time, and the experience and satisfaction of the patient are lower. Accordingly, an optimized hospital call management system is desired.
Accordingly, considering that in the process of registering and queuing for waiting for an actual patient, in order to accurately estimate waiting time, to reasonably plan and decide own journey based on the estimated time, for example, for a patient needing to hang several departments or numbers at the same time, the patient can choose which doctor to see first and then another doctor to improve experience and satisfaction of the patient. That is, since the condition and treatment of the treatment are different for each subject, such as assay, examination, start or no treatment, etc., the time spent by each subject at the time of the treatment is different, which is presented on the appointment information of the subject. However, since each subject to be treated has a different condition, the semantic understanding information in the treatment reservation information is also different, and the treatment reservation semantic understanding characteristics of each subject to be treated have an association relationship with respect to the waiting time, for example, when one patient leaves, the total waiting time needs to be recalculated. Therefore, in this process, the difficulty is how to dig out the related feature distribution information among the semantic understanding features of the appointment information of each to-be-diagnosed object, so as to accurately estimate the waiting time of the appointment number, so as to reasonably plan the own journey based on the estimated waiting time, and promote the experience and satisfaction of the patient.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. The development of deep learning and neural networks provides new solutions and schemes for mining the correlation feature distribution information among semantic understanding features of the appointment information of each to-be-diagnosed object.
Specifically, in the technical scheme of the present application, first, the appointment information of a plurality of objects to be treated arranged in front of the target object to be treated is acquired.
Specifically, in the embodiment of the present application, the appointment information semantic understanding module 120 is configured to obtain a plurality of appointment information semantic understanding feature vectors by respectively passing the appointment information of the respective objects to be treated through a context encoder including a word embedding layer. Then, considering that the appointment information of each to-be-diagnosed object is composed of each word, and semantic association characteristic information of context exists among each word, encoding each appointment information of each to-be-diagnosed object through a context encoder comprising a word embedding layer so as to extract relevant characteristic information of disorder context semantic understanding about each to-be-diagnosed object in each appointment information of each to-be-diagnosed object, namely hidden characteristic information of the diagnosis time of each to-be-diagnosed object, thereby obtaining a plurality of appointment information semantic understanding characteristic vectors.
Fig. 3 is a block diagram of the subscription information semantic understanding module in the hospital number calling management system according to the embodiment of the present application, as shown in fig. 3, the subscription information semantic understanding module 120 includes: a word segmentation unit 121, configured to perform word segmentation processing on the diagnosis appointment information of the respective subjects to be diagnosed, so as to convert the diagnosis appointment information of the respective subjects to be diagnosed into a word sequence composed of a plurality of words; an embedded encoding unit 122, configured to map each word in the word sequence to a word vector using an embedding layer of the context encoder including a word embedding layer to obtain a sequence of word vectors; and a context encoding unit 123, configured to perform global-based context semantic encoding on the sequence of word vectors using the converter of the context encoder including the word embedding layer to obtain the plurality of diagnosis appointment information semantic understanding feature vectors.
Further, fig. 4 is a block diagram of the context encoding unit in the hospital number calling management system according to the embodiment of the present application, and as shown in fig. 4, the context encoding unit 123 includes: a query vector construction subunit 1231, configured to one-dimensionally arrange the sequence of word vectors to obtain a global word feature vector; a self-attention subunit 1232 configured to calculate products between the global word feature vector and transpose vectors of respective word vectors in the sequence of word vectors to obtain a plurality of self-attention association matrices; a normalization subunit 1233, configured to perform normalization processing on each of the plurality of self-attention correlation matrices to obtain a plurality of normalized self-attention correlation matrices; a attention calculating subunit 1234 configured to obtain a plurality of probability values from each normalized self-attention correlation matrix in the plurality of normalized self-attention correlation matrices by using a Softmax classification function; and an attention applying subunit 1235 configured to weight each word vector in the sequence of word vectors with each probability value in the plurality of probability values as a weight to obtain the plurality of diagnosis appointment information semantic understanding feature vectors.
The context encoder aims to mine for hidden patterns between contexts in the word sequence, optionally the encoder comprises: CNN (Convolutional Neural Network ), recurrent NN (RecursiveNeural Network, recurrent neural network), language Model (Language Model), and the like. The CNN-based method has a better extraction effect on local features, but has a poor effect on Long-Term Dependency (Long-Term Dependency) problems in sentences, so Bi-LSTM (Long Short-Term Memory) based encoders are widely used. The repetitive NN processes sentences as a tree structure rather than a sequence, has stronger representation capability in theory, but has the weaknesses of high sample marking difficulty, deep gradient disappearance, difficulty in parallel calculation and the like, so that the repetitive NN is less in practical application. The transducer has a network structure with wide application, has the characteristics of CNN and RNN, has a better extraction effect on global characteristics, and has a certain advantage in parallel calculation compared with RNN (RecurrentNeural Network ).
Specifically, in the embodiment of the present application, the appointment information semantic association module 130 is configured to arrange the plurality of appointment information semantic understanding feature vectors into one-dimensional feature vectors, and obtain the appointment information semantic association feature vectors through a dual-branch association feature extraction structure including a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model. Further, it is also considered that there is a correlation between the condition context semantic understanding characteristics of the individual subjects to be treated with respect to latency, that is, the time of treatment of the individual subjects to be treated who are ranked in front of the target subject together determines the time that the target subject needs to wait, and when one or more patients in front of the target subject leave, the overall latency also needs to be recalculated.
Therefore, in the technical scheme of the application, after the plurality of appointment information semantic understanding feature vectors are arranged into one-dimensional feature vectors, the appointment information semantic association feature vectors are obtained through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model. Particularly, the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model adopt one-dimensional convolutional kernels with different scales to perform feature mining on the one-dimensional feature vectors so as to extract multi-scale relevance feature distribution information among semantic understanding features of disorder waiting time of each object to be diagnosed, and therefore the reservation information semantic relevance feature vectors are obtained. In this way, the time sequence multi-scale associated characteristic information of the treatment waiting time of each treatment object can be extracted.
Fig. 5 is a block diagram of the subscription information semantic association module in the hospital number calling management system according to the embodiment of the present application, as shown in fig. 5, the subscription information semantic association module 130 includes: a first convolution unit 131, configured to perform, in forward transfer of layers, one-dimensional convolution kernel-based convolution processing, feature matrix-based mean pooling processing, and activation processing on the one-dimensional feature vector by using each layer of a first one-dimensional convolution neural network model of the dual-branch correlation feature extraction structure, where the first one-dimensional convolution neural network model has a one-dimensional convolution kernel of a first scale, to output, by a last layer of the first one-dimensional convolution neural network model, a first-scale reservation information feature vector; a second convolution unit 132, configured to perform, in forward transfer of layers, convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix, and activation processing on the one-dimensional feature vector, respectively, using each layer of a second one-dimensional convolution neural network model of the dual-branch correlation feature extraction structure, to output, by a last layer of the second one-dimensional convolution neural network model, a second-scale reservation information feature vector, where the second one-dimensional convolution neural network model has a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; and a concatenation unit 133, configured to concatenate the first-scale subscription information feature vector and the second-scale subscription information feature vector to obtain the subscription information semantic association feature vector.
It should be noted that the multi-scale neighborhood feature extraction module is essentially a deep neural network model based on deep learning, which is capable of fitting any function by a predetermined training strategy and has a higher feature extraction generalization capability compared to the conventional feature engineering.
The multi-scale neighborhood feature extraction module comprises a plurality of parallel one-dimensional convolution layers, wherein in the process of feature extraction by the multi-scale neighborhood feature extraction module, the plurality of parallel one-dimensional convolution layers respectively carry out one-dimensional convolution coding on input data by one-dimensional convolution check with different scales so as to capture local implicit features of a sequence.
Specifically, in the embodiment of the present application, the latency evaluation module 140 is configured to pass the subscription information semantic association feature vector through a decoder to obtain a decoded value, where the decoded value is used to represent the estimated latency. And then, carrying out decoding regression on the reservation information semantic association feature vector in a decoder to obtain a decoding value for representing the estimated waiting time. That is, decoding is performed according to the multi-scale correlation characteristics of the treatment waiting time of each to-be-treated object arranged in front of the target treatment object, so that the treatment waiting time of the target treatment object is evaluated, the target treatment object can reasonably plan own journey based on the estimated waiting time, and the experience degree and satisfaction of patients are improved.
In a specific example of the application, the decoder is used for carrying out decoding regression on the reservation information semantic association feature vector according to the following decoding formula to obtain the decoding value; wherein, the decoding formula is:V d representing the reservation information semantic association feature vector, Y representing the decoded value, W representing the weight matrix, B representing the bias vector,>representation matrixAnd (5) multiplying.
Further, the hospital number calling management system also comprises a training module for training the context encoder comprising the word embedding layer, the double-branch association feature extraction structure comprising the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model and the decoder. Fig. 6 is a block diagram of the training module in the hospital number calling management system according to the embodiment of the present application, as shown in fig. 6, the training module 150 includes: the training data acquisition module 151 is configured to acquire training data, where the training data includes training appointment information of a plurality of objects to be diagnosed, and a true value of the estimated waiting time; the training appointment information semantic understanding module 152 is configured to obtain a plurality of training appointment information semantic understanding feature vectors by passing the training appointment information of the respective objects to be treated through the context encoder including the word embedding layer; the training appointment information semantic association module 153 is configured to arrange the plurality of training appointment information semantic understanding feature vectors into training one-dimensional feature vectors, and obtain training appointment information semantic association feature vectors through the two-branch association feature extraction structure including the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model; the optimization module 154 is configured to perform feature distribution optimization on the training reservation information semantic association feature vector to obtain an optimized training reservation information semantic association feature vector; a decoding loss module 155, configured to pass the optimization training reservation information semantic association feature vector through the decoder to obtain a decoding loss function value; and a training module 156 for training the context encoder including the word embedding layer, the dual-branch correlation feature extraction structure including the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model, and the decoder based on the decoding loss function value and by back propagation of gradient descent.
In particular, in the technical solution of the present application, in order to improve the feature expression effect of the subscription information semantic association feature vector, it is preferable to directly concatenate the first sub-feature vector and the second sub-feature vector obtained by the dual-branch association feature extraction structure of the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model to obtain the subscription information semantic association feature vector, however, considering that the input of the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model is the direct cascade arrangement of a plurality of diagnosis subscription information semantic understanding feature vectors, this may make the continuity of the overall feature distribution of the subscription information semantic association feature vector worse, thereby affecting the training effect.
Based on this, the applicant of the present application performs a normal periodic re-parameterization on the subscription information semantic association feature vector, for example denoted as V Geng Beier (gummel), to obtain an optimized subscription information semantic association feature vector, for example denoted as V', specifically: performing Geng Beier normal periodic re-parameterization on the training reservation information semantic association feature vector by using the following optimization formula to obtain the optimization reservation information semantic association feature vector; wherein, the optimization formula is:
Wherein v is i The characteristic values of all positions of the training reservation information semantic association characteristic vector are represented, mu and sigma are respectively the mean value and the variance of a characteristic value set of all positions of the training reservation information semantic association characteristic vector, log represents a logarithmic function based on 2, arcsin(s) represents an arcsine function, arccos(s) represents an arccosine function, v i ' feature values representing respective positions of the optimization training subscription information semantic association feature vector.
Here, the Geng Beier normal periodic re-parameterization is performed by semantically associating the reservation information with the feature values V of the respective positions of the feature vector V i Is converted into the angular feature expression of the probability distribution thereof to introduce the random periodic distribution in the normal distribution of the feature value set based on the random periodic operation mode of Geng Beier (Gumbel) distribution so as to obtain the periodic continuous with the randomness of the original feature distributionAnd the continuous approximation can be carried out, so that the dynamic continuous wave capacity of back propagation is improved through the periodical re-parameterization of the features, the dynamic applicability of a convolution kernel in the training process is improved, and the influence of poor continuity of the overall feature distribution of the reservation information semantic association feature vector on the training effect is compensated. Therefore, the waiting time for the patient to visit the patient can be accurately estimated, so that the journey of the patient can be reasonably planned based on the estimated waiting time, and the experience and satisfaction of the patient are improved.
Fig. 7 is a block diagram of the decoding loss module in the hospital number calling management system according to the embodiment of the present application, as shown in fig. 7, the decoding loss module 155 includes: a training decoding unit 1551, configured to perform decoding regression on the optimized training reservation information semantic association feature vector by using the decoder according to the following formula to obtain a training decoding value; wherein, the formula is:wherein X is the semantic association feature vector of the optimized training reservation information, Y is the training decoding value, W is a weight matrix,/and a method for optimizing the training reservation information>Representing a matrix multiplication; and a loss function calculation unit 1552 for calculating a variance between the training decoded value and a true value of the estimated wait time in the training data as the decoding loss function value.
In summary, the hospital number calling management system 100 according to the embodiment of the present application is illustrated, which acquires the appointment information of a plurality of objects to be treated arranged in front of the target object to be treated; and mining the correlation characteristic distribution information among the semantic understanding characteristics of the appointment information of each object to be diagnosed by adopting an artificial intelligence technology based on deep learning, accurately estimating the waiting time of the appointment number based on the correlation characteristic distribution information, reasonably planning the own journey based on the estimated waiting time, and improving the experience degree and satisfaction of patients.
As described above, the hospital number calling management system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server or the like for hospital number calling management. In one example, the hospital number management system 100 according to embodiments of the present application may be integrated into the terminal device as a software module and/or hardware module. For example, the hospital number management system 100 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the hospital number management system 100 may also be one of a plurality of hardware modules of the terminal device.
Alternatively, in another example, the hospital number management system 100 and the terminal device may be separate devices, and the hospital number management system 100 may be connected to the terminal device through a wired and/or wireless network and transmit the interactive information in a contracted data format.
In one embodiment of the present application, fig. 8 is a flowchart of a hospital number calling management method according to an embodiment of the present application. As shown in fig. 8, a hospital number calling management method according to an embodiment of the present application includes: 210, acquiring appointment information of a plurality of objects to be treated arranged in front of a target object to be treated; 220, the appointment information of each to-be-treated object is respectively passed through a context encoder comprising a word embedding layer to obtain a plurality of appointment information semantic understanding feature vectors; 230, arranging the plurality of appointment information semantic understanding feature vectors into one-dimensional feature vectors, and obtaining appointment information semantic association feature vectors through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model; and, 240, passing the reservation information semantically-related feature vector through a decoder to obtain a decoded value, wherein the decoded value is used for representing the estimated waiting time.
Fig. 9 is a schematic diagram of a system architecture of a hospital number calling management method according to an embodiment of the present application. As shown in fig. 9, in the system architecture of the hospital number calling management method, first, the appointment information of a plurality of objects to be treated arranged in front of a target object to be treated is acquired; then, the appointment information of each to-be-diagnosed object is respectively passed through a context encoder comprising a word embedding layer to obtain a plurality of appointment information semantic understanding feature vectors; secondly, arranging the plurality of appointment information semantic understanding feature vectors into one-dimensional feature vectors, and obtaining appointment information semantic association feature vectors through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model; and finally, the reservation information semantic association feature vector is passed through a decoder to obtain a decoded value, wherein the decoded value is used for representing the estimated waiting time.
In a specific example, in the hospital number calling management method, the method for obtaining a plurality of diagnosis reservation information semantic understanding feature vectors by passing the diagnosis reservation information of each to-be-diagnosed object through a context encoder including a word embedding layer includes: performing word segmentation processing on the appointment information of each to-be-diagnosed object so as to convert the appointment information of each to-be-diagnosed object into a word sequence composed of a plurality of words; mapping each word in the word sequence to a word vector using an embedding layer of the context encoder including the word embedding layer to obtain a sequence of word vectors; and performing global-based context semantic coding on the sequence of word vectors using a translator of the context encoder including a word embedding layer to obtain the plurality of appointment information semantic understanding feature vectors.
In a specific example, in the hospital number calling management method, the step of performing global-based context semantic coding on the sequence of word vectors by using the converter of the context encoder including the word embedding layer to obtain the plurality of diagnosis appointment information semantic understanding feature vectors includes: one-dimensional arrangement is carried out on the sequence of the word vectors to obtain global word feature vectors; calculating the product between the global word feature vector and the transpose vector of each word vector in the sequence of word vectors to obtain a plurality of self-attention association matrices; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and weighting each word vector in the sequence of word vectors by taking each probability value in the plurality of probability values as a weight to obtain the plurality of appointment information semantic understanding feature vectors.
In a specific example, in the hospital number calling management method, after the plurality of appointment information semantic understanding feature vectors are arranged into one-dimensional feature vectors, the appointment information semantic association feature vectors are obtained through a dual-branch association feature extraction structure including a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model, and the method includes: each layer of a first one-dimensional convolutional neural network model of the double-branch correlation feature extraction structure is used for respectively carrying out one-dimensional convolutional kernel-based convolutional processing, feature matrix-based mean pooling processing and activating processing on the one-dimensional feature vector in forward transmission of the layer, so that the output of the last layer of the first one-dimensional convolutional neural network model is used as a first-scale reservation information feature vector, wherein the first one-dimensional convolutional neural network model is provided with a one-dimensional convolutional kernel of a first scale; performing one-dimensional convolution kernel-based convolution processing, feature matrix-based averaging pooling processing and activation processing on the one-dimensional feature vector in forward transfer of layers by using each layer of a second one-dimensional convolution neural network model of the dual-branch correlation feature extraction structure to obtain a second-scale reservation information feature vector by using the output of the last layer of the second one-dimensional convolution neural network model, wherein the second one-dimensional convolution neural network model has one-dimensional convolution kernels of a second scale, and the first scale is different from the second scale; and cascading the first-scale reservation information feature vector and the second-scale reservation information feature vector to obtain the reservation information semantic association feature vector.
In a specific example, in the hospital number calling management method, training is further performed on the context encoder including the word embedding layer, the two-branch correlation feature extraction structure including the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model, and the decoder.
In a specific example, in the hospital number calling management method, training the context encoder including the word embedding layer, the two-branch correlation feature extraction structure including the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model, and the decoder includes: acquiring training data, wherein the training data comprises training diagnosis reservation information of a plurality of objects to be diagnosed and the estimated real value of the waiting time; respectively passing the training appointment information of each object to be diagnosed through the context encoder comprising the word embedding layer to obtain a plurality of training appointment information semantic understanding feature vectors; arranging the training appointment information semantic understanding feature vectors into training one-dimensional feature vectors, and then obtaining training appointment information semantic association feature vectors through the double-branch association feature extraction structure comprising the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model; performing feature distribution optimization on the training reservation information semantic association feature vectors to obtain optimized training reservation information semantic association feature vectors; the optimized training reservation information semantic association feature vector passes through the decoder to obtain a decoding loss function value; and training the context encoder including the word embedding layer, the dual-branch correlation feature extraction structure including the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model, and the decoder based on the decoding loss function value and by back propagation of gradient descent.
In a specific example, in the hospital number calling management method, performing feature distribution optimization on the training reservation information semantic association feature vector to obtain an optimized training reservation information semantic association feature vector includes: performing Geng Beier normal periodic re-parameterization on the training reservation information semantic association feature vector by using the following optimization formula to obtain the optimization reservation information semantic association feature vector; wherein, the optimization formula is:
wherein v is i The characteristic values of all positions of the training reservation information semantic association characteristic vector are represented, mu and sigma are respectively the mean value and the variance of a characteristic value set of all positions of the training reservation information semantic association characteristic vector, log represents a logarithmic function based on 2, arcsin(s) represents an arcsine function, arccos(s) represents an arccosine function, v i ' feature values representing respective positions of the optimization training subscription information semantic association feature vector.
In a specific example, in the hospital number calling management method, the step of passing the optimized training reservation information semantic association feature vector through the decoder to obtain a decoding loss function value includes: performing decoding regression on the optimized training reservation information semantic association feature vector by using the decoder according to the following formula to obtain a training decoding value; wherein, the formula is: Wherein X is the semantic association feature vector of the optimized training reservation information, Y is the training decoding value, W is a weight matrix,/and a method for optimizing the training reservation information>Representing a matrix multiplication; and calculating a variance between the training decoded value and a true value of the estimated wait time in the training data as the decoding loss function value.
It will be appreciated by those skilled in the art that the specific operations of the respective steps in the above-described hospital number management method have been described in detail in the above description of the hospital number management system with reference to fig. 1 to 7, and thus, repetitive descriptions thereof will be omitted.
The present application also provides a computer program product comprising instructions which, when executed, cause an apparatus to perform operations corresponding to the above-described methods.
In one embodiment of the present application, there is also provided a computer readable storage medium storing a computer program for executing the above-described method.
It should be appreciated that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the forms of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects may be utilized. Furthermore, the computer program product may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
Methods, systems, and computer program products of embodiments of the present application are described in terms of flow diagrams and/or block diagrams. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not limiting, and these advantages, benefits, effects, etc. are not to be considered as necessarily possessed by the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not intended to be limited to the details disclosed herein as such.
The block diagrams of the devices, apparatuses, devices, systems referred to in this application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.
It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent to the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Finally, it is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or terminal device comprising the element.
The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.

Claims (10)

1. A hospital number calling management system, comprising:
the system comprises a diagnosis appointment information data acquisition module, a diagnosis appointment information acquisition module and a diagnosis appointment information processing module, wherein the diagnosis appointment information acquisition module is used for acquiring diagnosis appointment information of a plurality of objects to be diagnosed arranged in front of a target object to be diagnosed;
the appointment information semantic understanding module is used for enabling the appointment information of each object to be diagnosed to pass through a context encoder comprising a word embedding layer to obtain a plurality of appointment information semantic understanding feature vectors;
the appointment information semantic association module is used for arranging the appointment information semantic understanding feature vectors for the doctor for a plurality of doctor-seeing to be one-dimensional feature vectors and obtaining appointment information semantic association feature vectors through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model; and
and the waiting time evaluation module is used for enabling the reservation information semantic association feature vector to pass through a decoder to obtain a decoding value, wherein the decoding value is used for representing the estimated waiting time.
2. The hospital number calling management system according to claim 1, wherein the reservation information semantic understanding module comprises:
the word segmentation unit is used for carrying out word segmentation processing on the appointment information of each to-be-diagnosed object so as to convert the appointment information of each to-be-diagnosed object into a word sequence consisting of a plurality of words;
an embedded encoding unit, configured to map each word in the word sequence to a word vector using an embedding layer of the context encoder including a word embedding layer to obtain a sequence of word vectors; and
and the context coding unit is used for carrying out global-based context semantic coding on the sequence of the word vectors by using the converter of the context coder comprising the word embedding layer so as to obtain the plurality of diagnosis appointment information semantic understanding feature vectors.
3. The hospital number calling management system according to claim 2, wherein said context encoding unit comprises:
a query vector construction subunit, configured to perform one-dimensional arrangement on the sequence of word vectors to obtain a global word feature vector;
a self-attention subunit, configured to calculate a product between the global word feature vector and a transpose vector of each word vector in the sequence of word vectors to obtain a plurality of self-attention association matrices;
The normalization subunit is used for respectively performing normalization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of normalized self-attention correlation matrices;
the attention calculating subunit is used for obtaining a plurality of probability values through a Softmax classification function by each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and
and the attention applying subunit is used for weighting each word vector in the word vector sequence by taking each probability value in the plurality of probability values as a weight so as to obtain the plurality of diagnosis appointment information semantic understanding feature vectors.
4. The hospital number calling management system according to claim 3, wherein said reservation information semantic association module comprises:
the first convolution unit is used for respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and activation processing on the one-dimensional feature vector in forward transfer of layers by using each layer of a first one-dimensional convolution neural network model of the double-branch correlation feature extraction structure to take the output of the last layer of the first one-dimensional convolution neural network model as a first-scale reservation information feature vector, wherein the first one-dimensional convolution neural network model is provided with a one-dimensional convolution kernel of a first scale;
A second convolution unit, configured to perform convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix, and activation processing on the one-dimensional feature vector in forward transfer of layers of a second one-dimensional convolution neural network model using the two-branch correlation feature extraction structure, so that an output of a last layer of the second one-dimensional convolution neural network model is a second-scale reservation information feature vector, where the second one-dimensional convolution neural network model has a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; and
and the cascading unit is used for cascading the first-scale reservation information feature vector and the second-scale reservation information feature vector to obtain the reservation information semantic association feature vector.
5. The hospital number management system according to claim 4, further comprising a training module for training the context encoder including the word embedding layer, the two-branch correlation feature extraction structure including the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model, and the decoder.
6. The hospital number management system according to claim 5, wherein the training module comprises:
the training data acquisition module is used for acquiring training data, wherein the training data comprises training diagnosis reservation information of a plurality of objects to be diagnosed and the estimated real value of the waiting time;
the training appointment information semantic understanding module is used for enabling the training appointment information of each object to be diagnosed to pass through the context encoder comprising the word embedding layer to obtain a plurality of training appointment information semantic understanding feature vectors;
the training appointment information semantic association module is used for arranging the plurality of training appointment information semantic understanding feature vectors into training one-dimensional feature vectors and obtaining training appointment information semantic association feature vectors through the double-branch association feature extraction structure comprising the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model;
the optimization module is used for carrying out feature distribution optimization on the training reservation information semantic association feature vectors so as to obtain optimization training reservation information semantic association feature vectors;
the decoding loss module is used for enabling the optimization training reservation information semantic association feature vectors to pass through the decoder so as to obtain decoding loss function values; and
The training module is used for training the context encoder comprising the word embedding layer, the double-branch associated feature extraction structure comprising the first one-dimensional convolutional neural network model and the second one-dimensional convolutional neural network model and the decoder based on the decoding loss function value and through back propagation of gradient descent.
7. The hospital call management system according to claim 6, wherein said optimization module is configured to: performing Geng Beier normal periodic re-parameterization on the training reservation information semantic association feature vector by using the following optimization formula to obtain the optimization reservation information semantic association feature vector;
wherein, the optimization formula is:
wherein v is i The characteristic values of all positions of the training reservation information semantic association characteristic vector are represented, mu and sigma are respectively the mean value and the variance of a characteristic value set of all positions of the training reservation information semantic association characteristic vector, log represents a logarithmic function based on 2, arcsin(s) represents an arcsine function, arccos(s) represents an arccosine function, v i ' feature values representing respective positions of the optimization training subscription information semantic association feature vector.
8. The hospital number management system according to claim 7, wherein said decode loss module comprises:
the training decoding unit is used for carrying out decoding regression on the semantic association feature vectors of the optimized training reservation information by using the decoder according to the following formula so as to obtain training decoding values; wherein, the formula is:wherein X is the semantic association feature vector of the optimized training reservation information, Y is the training decoding value, W is a weight matrix,/and a method for optimizing the training reservation information>Representing a matrix multiplication; and
and a loss function calculation unit for calculating a variance between the training decoded value and a true value of the estimated wait time in the training data as the decoding loss function value.
9. A hospital number calling management method, comprising:
acquiring appointment information of a plurality of objects to be treated, which are arranged in front of a target object to be treated;
the appointment information of each object to be diagnosed is respectively passed through a context encoder comprising a word embedding layer to obtain a plurality of appointment information semantic understanding feature vectors;
arranging the plurality of appointment information semantic understanding feature vectors into one-dimensional feature vectors, and obtaining appointment information semantic association feature vectors through a double-branch association feature extraction structure comprising a first one-dimensional convolutional neural network model and a second one-dimensional convolutional neural network model; and
And the reservation information semantic association feature vector is passed through a decoder to obtain a decoding value, wherein the decoding value is used for representing the estimated waiting time.
10. The hospital number calling management method according to claim 9, wherein passing the appointment information of each object to be treated through a context encoder including a word embedding layer to obtain a plurality of appointment information semantic understanding feature vectors, respectively, comprises:
performing word segmentation processing on the appointment information of each to-be-diagnosed object so as to convert the appointment information of each to-be-diagnosed object into a word sequence composed of a plurality of words;
mapping each word in the word sequence to a word vector using an embedding layer of the context encoder including the word embedding layer to obtain a sequence of word vectors; and
and performing global-based context semantic coding on the sequence of word vectors by using a converter of the context encoder comprising the word embedding layer to obtain the plurality of appointment information semantic understanding feature vectors.
CN202310430032.4A 2023-04-21 2023-04-21 Hospital number calling management system and method thereof Pending CN116469534A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310430032.4A CN116469534A (en) 2023-04-21 2023-04-21 Hospital number calling management system and method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310430032.4A CN116469534A (en) 2023-04-21 2023-04-21 Hospital number calling management system and method thereof

Publications (1)

Publication Number Publication Date
CN116469534A true CN116469534A (en) 2023-07-21

Family

ID=87181950

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310430032.4A Pending CN116469534A (en) 2023-04-21 2023-04-21 Hospital number calling management system and method thereof

Country Status (1)

Country Link
CN (1) CN116469534A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116823193A (en) * 2023-08-31 2023-09-29 深圳市永迦电子科技有限公司 Intelligent manufacturing flow management system based on big data

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116823193A (en) * 2023-08-31 2023-09-29 深圳市永迦电子科技有限公司 Intelligent manufacturing flow management system based on big data
CN116823193B (en) * 2023-08-31 2023-11-03 深圳市永迦电子科技有限公司 Intelligent manufacturing flow management system based on big data

Similar Documents

Publication Publication Date Title
CN110516161B (en) Recommendation method and device
Hassantabar et al. CovidDeep: SARS-CoV-2/COVID-19 test based on wearable medical sensors and efficient neural networks
CN106295186B (en) A kind of system of the aided disease diagnosis based on intelligent inference
CN109741806B (en) Auxiliary generation method and device for medical image diagnosis report
KR20170061222A (en) The method for prediction health data value through generation of health data pattern and the apparatus thereof
CN109378066A (en) A kind of control method and control device for realizing disease forecasting based on feature vector
CN110111886A (en) A kind of intelligent interrogation system and method based on XGBoost disease forecasting
CN109949929A (en) A kind of assistant diagnosis system based on the extensive case history of deep learning
CN111984771A (en) Automatic inquiry system based on intelligent conversation
WO2020224433A1 (en) Target object attribute prediction method based on machine learning and related device
CN111666477A (en) Data processing method and device, intelligent equipment and medium
US10847261B1 (en) Methods and systems for prioritizing comprehensive diagnoses
WO2023178971A1 (en) Internet registration method, apparatus and device for seeking medical advice, and storage medium
EP4156202A1 (en) Method and system for predicting needs of patient for hospital resources
CN115497616A (en) Method, system, equipment and storage medium for aid decision making of infectious diseases
CN113160993A (en) Health intervention system, server and health management system
CN111696661A (en) Patient clustering model construction method, patient clustering method and related equipment
CN116469534A (en) Hospital number calling management system and method thereof
CN112908452A (en) Event data modeling
Yan et al. Kidney stone detection using an optimized Deep Believe network by fractional coronavirus herd immunity optimizer
CN110088748B (en) Question generation method and device, question inquiry system and computer readable storage medium
CN111159369A (en) Multi-round intelligent inquiry method and device and computer readable storage medium
CN112259243B (en) Single-disease clinical path construction method and system based on bidirectional GRUs
CN117034142A (en) Unbalanced medical data missing value filling method and system
CN117253629A (en) Medical guide information pushing method, device, equipment, medium and computer program product

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination